On July 14, the Kyligence User Conference with the theme of "Unleashing Digital Intelligence Productivity" was successfully held in Shanghai. A member of Kyligence's latest product family: the preview version of Kyligence Copilot, an AI digital assistant, was released at the conference. At present, Kyligence Copilot has also opened the application test, and you can visit the application channel at the end of the article.
Click the link to view the video to see what Copilot is capable of. Kyligence, praise 98 We can see that Copilot can understand business problems well, convert them into corresponding analysis and suggestions, and feed back the results to users through natural language, which greatly reduces the threshold for users to use data to drive business decisions. What is the core here?
- Model pre-training based on a large number of parameters?
- Model fine-tuning (Fine-tune)?
- Complicated tip works?
Neither is quite accurate. Kyligence Copilot is an intelligent application based on large language model services (LLM Services) and indicator data services (Metrics & Data Services), which packs esoteric Data + AI technology into product functions that business personnel can use with a low threshold.
Kyligence Copilot architecture diagram
Indicators make AI understand business better
Kyligence Copilot is based on the intelligent capabilities provided by Kyligence Zen's one-stop indicator platform, which determines that a unique feature of Kyligence Copilot is that all data insights and analysis suggestions are generated based on indicators. Indicators are the projection of business in the technical space. AI can understand the business context more deeply by understanding the definition of indicators and reading the results of indicators, and then give more professional answers to business questions raised by users.
The metrics platform provides AI with a limited business context, allowing AI to generate content within a limited range that meets user expectations. For example, AI is a generalist with high comprehensive quality, and the indicator system maintained by the enterprise through the indicator platform provides AI with professional and comprehensive business data, so that AI can solve the business problems represented by these business data. in-depth thinking. In addition, the indicator caliber is managed uniformly through the indicator platform to prevent AI from accepting ambiguous business data and improve the accuracy of AI insights.
Bridge between big model and big data
Kyligence Copilot itself is not a large model, but an application based on large model technologies such as OpenAI. Kyligence Copilot provides functions such as Metrics Discovery & Deep Dive, KPI Evaluation & Suggestion, Workflow & Data Products, etc. The core behind it is generated with the help of large model services Insights and recommendations based on indicator data, and feedback to users.
To give a specific example, when Kyligence Copilot answers user questions, the execution logic behind it is roughly divided into four steps:
- The first step is to review the question to see whether the question asked is legal, compliant, logical, etc. This is the first dialogue with the language model.
- The second step is instruction understanding , which is to map user requests into specific actions of an indicator platform. This step is usually difficult because users express various business needs.
- The third step is to execute the instructions on the indicator platform , which may be attribution analysis, target Kanban, etc. What is obtained in this step is only data.
- The fourth step is to use a language model to interpret the data and charts in natural language, and to feed back key insights to users.
If a worker wants to do a good job, he must first sharpen his tools. You can see that the ability of OpenAI ChatGPT is quite good, how about the performance of other large models in terms of understanding and performance? We have also done some tests and research for you. We have used an open source large language model instead of ChatGPT to make some preliminary attempts. For example, LLaMA 13B and Falcon 40B can probably reach about 70% of ChatGPT 3.5 in terms of test command understanding ability. Capabilities, which can be understood as the minimum capability range that enterprises hope to implement Copilot, a large language model indicator, locally.
How to control costs?
The biggest difference between Kyligence Copilot and traditional data tools is that traditional tools (even Excel) are professional tools, which determines that only some people in the enterprise may be able to use them, while Kyligence Copilot will be used by everyone in the enterprise. We can expect that the load of the Kyligence Copilot analysis engine may increase by hundreds or even thousands of times, which creates stringent requirements for cost optimization. And this is also the direction that Kyligence has been working on, that is, how to use an ultra-high concurrent OLAP engine technology to support a hundred times the load.
Since Kyligence released the AI-enhanced engine in 2019, Kyligence has been continuously optimizing the performance and concurrency of multi-dimensional analysis scenarios through AI + pre-computing technology, while reducing costs. In addition, we continue to improve the performance of the computing engine. The vectorized Spark engine technology we developed, Kyligence Turbo, has more than doubled the speed of the standard Spark, which can help enterprises save about 50% of computing power and costs.
Trials of Kyligence Zen and Kyligence Copilot are now open, and you are welcome to click the link to apply for a trial.
About Kyligence
Founded in 2016 by the founding team of Apache Kylin, Kyligence is a leading provider of big data analysis and indicator platforms, providing enterprise-level OLAP (multidimensional analysis) product Kyligence Enterprise and one-stop indicator platform Kyligence Zen for users Provide enterprise-level business analysis capabilities, decision support systems and various data-driven industry solutions.
Kyligence has served many customers in banking, securities, insurance, manufacturing, retail, medical and other industries in China, the United States, Europe and Asia Pacific, including China Construction Bank, Ping An Bank, Shanghai Pudong Development Bank, Bank of Beijing, Bank of Ningbo, Pacific Insurance, China UnionPay, SAIC, Changan Automobile, Starbucks, Anta, Li Ning, AstraZeneca, UBS, MetLife and other world-renowned companies, and reached global partnerships with Microsoft, Amazon Cloud Technology, Huawei, Ernst & Young, Deloitte, etc. Kyligence has received multiple investments from institutions such as Redpoint, Broadband Capital, Shunwei Capital, Eight Roads Capital, Coatue, SPDB International, CICC Capital, Gopher Assets, and Guofang Capital.